This article was downloaded by: [University of California Santa Cruz]On: 14 May 2014, At: 15:05Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,37-41 Mortimer Street, London W1T 3JH, UK

Transactions of the American Fisheries SocietyPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/utaf20

Contemporary Population Structure in KlamathRiver Basin Chinook Salmon Revealed by Analysis ofMicrosatellite Genetic DataAndrew P. Kinziger a , Michael Hellmair a , David G. Hankin a & John Carlos Garza ba Department of Fisheries Biology , Humboldt State University , One Harpst Street, Arcata ,California , 95521 , USAb Southwest Fisheries Science Center , Fisheries Ecology Division , 110 Shaffer Road, SantaCruz , California , 95060 , USAPublished online: 29 Aug 2013.

To cite this article: Andrew P. Kinziger , Michael Hellmair , David G. Hankin & John Carlos Garza (2013) ContemporaryPopulation Structure in Klamath River Basin Chinook Salmon Revealed by Analysis of Microsatellite Genetic Data, Transactionsof the American Fisheries Society, 142:5, 1347-1357, DOI: 10.1080/00028487.2013.806351

To link to this article: http://dx.doi.org/10.1080/00028487.2013.806351

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) containedin the publications on our platform. However, Taylor & Francis, our agents, and our licensors make norepresentations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of theContent. Any opinions and views expressed in this publication are the opinions and views of the authors, andare not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon andshould be independently verified with primary sources of information. Taylor and Francis shall not be liable forany losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoeveror howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use ofthe Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in anyform to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Transactions of the American Fisheries Society 142:1347–1357, 2013C© American Fisheries Society 2013ISSN: 0002-8487 print / 1548-8659 onlineDOI: 10.1080/00028487.2013.806351

ARTICLE

Contemporary Population Structure in Klamath River BasinChinook Salmon Revealed by Analysis of MicrosatelliteGenetic Data

Andrew P. Kinziger,* Michael Hellmair, and David G. HankinDepartment of Fisheries Biology, Humboldt State University, One Harpst Street, Arcata,California 95521, USA

John Carlos GarzaSouthwest Fisheries Science Center, Fisheries Ecology Division, 110 Shaffer Road, Santa Cruz,California 95060, USA

AbstractChinook Salmon Oncorhynchus tshawytscha exhibit substantial population genetic structure at multiple scales.

Although geography is generally more important than life history, particularly migration and run timing, for de-scribing genetic structure in Chinook Salmon, there are several exceptions to this general pattern, and hatcherysupplementation has altered natural genetic structure in some areas. Given that genetic structure of Chinook Salmonis often basin-specific, we assessed genetic variation of 27 microsatellite loci in geographically and temporally distinctnatural populations and hatchery stocks in the Klamath River basin, California. Multiple analyses support recogni-tion of three major genetic lineages from separate geographic regions in the Klamath River basin: the lower basin, theKlamath River, and the Trinity River. The lower basin group was sharply distinct, but populations in the Klamath andTrinity river lineages were connected by processes that can be described by a one-dimensional, linear, stepping-stonemodel where gene exchange occurred primarily, but not exclusively, between adjacent populations. Genetic structureby migration timing was also evident, although divergences among populations that differed by migration timing onlywere fewer than those observed between geographic regions. Distinct run-timing ecotypes in the Klamath River basinthus appear to have evolved independently through a process of parallel evolution. Introgressive pressure from thehatchery stocks into natural populations was attenuated by distance from the hatchery, but comparison of historicalpopulation genetic structure to contemporary patterns would be needed to fully evaluate the extent to which hatcherystocks may have altered natural genetic structure.

Chinook Salmon Oncorhynchus tshawytscha exhibits sub-stantial population genetic structure across its geographic range(Waples et al. 2004; Beacham et al. 2006). Population structurein Chinook Salmon is directly related to their strong propensityto home to natal spawning grounds with high accuracy. Natalhoming ultimately results in large genetic differentiation be-tween geographically separated river drainages, presumably dueto relatively low straying between basins (Waples et al. 2004;Beacham et al. 2006). Lower genetic divergence among pop-ulations from tributaries within the same drainage presumably

*Corresponding author: andrew.kinziger@humboldt.eduReceived December 20, 2012; accepted May 14, 2013Published online August 29, 2013

results from straying being more common. Chinook Salmon en-ter freshwater during almost every month of the year in largerbasins (Healey 1991) and Chinook Salmon populations often ex-hibit temporal genetic structure according to migration and runtiming, and different run times are frequently genetically diver-gent from one another (Waples et al. 2004). However, ChinookSalmon stocks with temporally distinct migration timing gener-ally exhibit less genetic differentiation than those stocks orig-inating from separate geographic regions (Waples et al. 2004;Moran et al. 2013). These findings support the hypothesis that

1347

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 15:

05 1

4 M

ay 2

014

1348 KINZIGER ET AL.

different migration times in Chinook Salmon have evolved re-peatedly and independently at different geographic locations,probably due to divergent selection on run timing (Waples et al.2004; Moran et al. 2013). However, there are exceptions to thesegeneral patterns. For example, Columbia River basin ChinookSalmon are primarily structured by life history (e.g., run timing,yearling and subyearling out-migrant types: Narum et al. 2010;Moran et al. 2013) and genetic structure of Chinook Salmonwithin the California Central Valley is more coincident with runtiming than with geography (Banks et al. 2000; Garza et al.2008).

Hatchery practices can also substantially alter the geneticstructure of Chinook Salmon within river basins. For example,off-site release of juvenile hatchery fish in the California CentralValley has resulted in genetic homogenization and the disappear-ance of population structure among fall-run Chinook Salmonpopulations in different tributaries (Williamson and May 2005).However, the effects of hatchery stocking of salmonids can behighly variable. In some cases, genetic structure has been to-tally disrupted, whereas, in others, introduced stocks do notappear to substantially contribute to the recipient population

(Hansen et al. 2001, 2009; Martinez et al. 2001; Hansen 2002;Williamson and May 2005; Finnegan and Stevens 2008; Matalaet al. 2012). Thus, understanding the genetic structure of Chi-nook Salmon within a particular river basin and how it is hasbeen affected by hatchery supplementation requires fine-scalesampling and genetic analysis of populations in that basin.

The Klamath River basin in California is one of the largestproducers of Chinook Salmon in western North America. Twolarge-scale hatchery programs were initiated in the basin in theearly 1960s to mitigate for habitat loss caused by constructionof impassable dams. Iron Gate Hatchery is at the currentupstream limit of anadromy on the main-stem Klamath River,and Trinity River Hatchery is at the current upstream limit of itslargest tributary (Figure 1). Iron Gate Hatchery raises fall-runChinook Salmon only, whereas Trinity River Hatchery raisesboth fall-run and spring-run fish. The spring-run salmon thathistorically spawned upstream from Iron Gate Hatchery arepresumed to have been extirpated (Snyder 1931; Hamilton et al.2005). Currently, hatchery releases number about 10 millionjuvenile Chinook Salmon annually (both hatcheries combined;Hamilton et al. 2011) and hatchery fish contribute substantially

FIGURE 1. The Klamath River basin, California–Oregon, depicting major tributaries, hatcheries, and dams. Iron Gate and Lewiston dams are barriers toanadromy and represent the extent of upstream migration in the Klamath and Trinity rivers, respectively. The arrow indicates the confluence of the Klamath andTrinity rivers and the division between the Upper Klamath and Trinity Rivers ESU (above) and the Southern Oregon and Northern California Coast ESU (below).

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 15:

05 1

4 M

ay 2

014

KLAMATH RIVER BASIN CHINOOK SALMON GENETIC STRUCTURE 1349

to ocean and in-river tribal, commercial, and recreationalharvest. While the hatchery stocks are derived from localsources, eggs have been exchanged between the hatcheries andhatchery fish have been released off-site in various KlamathRiver basin tributaries. The data in Supplementary Tables 1–3(in the online version of this article) are illustrative of thetypes of hatchery releases that have occurred in the KlamathRiver basin, but are not comprehensive. Overall, the extentof these activities for Klamath River Chinook Salmon is lessthan for most other basins with substantial hatchery production(Bjorkstedt et al. 2005; see also Spence et al. 2011)

Several Klamath River tributaries contain naturally spawningChinook Salmon stocks that are presumably outside of the in-fluence of hatchery stocking, due to large geographic separationfrom both Iron Gate and Trinity River hatcheries. These includethe Scott, Salmon, and South Fork Trinity rivers and smallertributaries in the lower basin (Figure 1). Two stocks from thesenatural areas, spring-run fish from the South Fork Trinity Riverand Salmon River, occur in low abundance and are of conser-vation concern. The South Fork Trinity River Chinook Salmonpopulation currently consists of a few hundred returning adults,down from >11,000 in 1964, and the Salmon River popula-tion is also at low abundance, estimated at 250–1,400 returningadults per year.

Klamath River Chinook Salmon have been divided into twoevolutionarily significant units (ESU) separated by the conflu-ence of the Klamath and Trinity rivers, based upon analysisof genetic and other data (Figure 1; Gall et al. 1992; Myerset al. 1998; Waples et al. 2004; Beacham et al. 2006; Seeb et al.2007). Populations downstream from the confluence are part ofthe Southern Oregon and Northern California Coast ESU, whichincludes additional stocks north of the Klamath River basin, in-cluding the Rogue River in Oregon. Populations upstream from

the confluence are assigned to the Upper Klamath and TrinityRivers ESU, which includes both spring- and fall-run stocks.The extent of genetic differentiation between spring- and fall-run stocks from the Trinity River Hatchery has previously beenshown to be low (Kinziger et al. 2008).

The objective of the current study was to determine within-basin genetic population structure of Chinook Salmon fromthe Klamath River basin and evaluate the effects of decadesof large-scale hatchery production on this population structure.We analyzed genotypic data from 27 highly informative mi-crosatellite loci for 790 fish from 10 natural populations andall three hatchery stocks (Table 1). Our study includes rep-resentatives of all major Chinook Salmon populations in theKlamath River basin comprising the vast majority of all naturalChinook Salmon production in the basin, including spring-runpopulations at very low levels of abundance. We provide thefirst comprehensive evaluation of Chinook Salmon populationstructure in the Klamath–Trinity River and establish a baselinefor evaluation of future trends in the basin.

METHODSSample collections.—Tissues (fins, scales, or both) were

collected during carcass surveys, operation of weirs, hatcheryspawning, or by electrofishing and placed into 95% ethanol,dried on blotter paper, or frozen until DNA extraction. All col-lections were from adults except those from Iron Gate Hatch-ery (IGH), Terwer Creek (TC), and Blue Creek (BC) (Table 1),which were from juveniles. Juvenile collections were spaced outtemporally to minimize problems associated with family sam-pling (e.g., Allendorf and Phelps 1981; Hansen et al. 1997). Toassess the extent of family sampling in our juvenile collections,we estimated pairwise relatedness among individuals within

TABLE 1. Population, run time, population abbreviation (ID), sample size (n), year of field collections, rarified private alleles (Ap), rarified allelic richness(AR), allelic richness (A), observed heterozygosity (Ho), expected heterozygosity (He), and proportion of membership in each of three Chinook Salmon populationclusters (lower basin, Klamath, and Trinity) inferred using STRUCTURE.

LowerPopulation Run time ID n Year Ap AR A Ho He basin Klamath Trinity

Iron Gate Hatchery Fall IGH 104 2006 0.3 9.7 15.0 0.74 0.74 0.06 0.88 0.06Bogus Creek Fall BOG 32 2006 0.3 9.4 10.3 0.72 0.74 0.08 0.84 0.08Shasta River Fall SHST 31 2002 0.3 9.8 11.3 0.73 0.74 0.08 0.85 0.07Scott River Fall SCOT 64 2002 0.4 10.7 15.1 0.75 0.76 0.18 0.64 0.19Salmon River Spring SRS 94 1997, 2006 0.1 10.0 13.3 0.73 0.75 0.17 0.22 0.61Salmon River Fall SRF 52 2002, 2006 0.3 10.9 16.7 0.75 0.77 0.24 0.28 0.48Trinity River Hatchery Spring TRHS 133 1992, 2004 0.2 9.5 15.3 0.72 0.74 0.06 0.06 0.88Trinity River Hatchery Fall TRHF 124 1992, 2004 0.1 9.6 15.0 0.74 0.74 0.06 0.07 0.87South Fork Trinity River Spring SFTS 7 1993 0.15 0.08 0.77South Fork Trinity River Fall SFTF 19 1993 0.3 9.7 9.7 0.69 0.74 0.22 0.20 0.58Horse Linto Creek Fall HLC 38 1997 0.4 10.5 12.4 0.75 0.77 0.81 0.08 0.11Blue Creek Fall BC 69 2008 0.7 12.1 17.5 0.79 0.80 0.90 0.05 0.05Terwer Creek Fall TC 23 2008 0.7 10.2 10.7 0.76 0.76 0.93 0.03 0.04

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 15:

05 1

4 M

ay 2

014

1350 KINZIGER ET AL.

each collection using ML-RELATE (Kalinowski et al. 2006). In-clusion of hybrids between spring- and fall-run salmon returningto Trinity River Hatchery (Kinziger et al. 2008) was minimizedby selecting samples from the early portion of the spring-runspawning period and the end of the fall-run spawning period.

Molecular methods.—The DNA was extracted using thePromega Wizard SV96 Genomic DNA Purification System, Qi-agen DNeasy spin columns using a Qiagen 3000 BioRobot, orwith a Chelex Resin protocol. A total of 28 microsatellite lociwere assayed (Supplementary Table 4). Sixteen loci were geno-typed using a Beckman Coulter CEQ 8000 Genetic AnalysisSystem and 12 loci, which have been standardized for use withChinook Salmon (Seeb et al. 2007), were genotyped with anApplied Biosystems 3730 genetic analyzer. Polymerase chainreaction volumes and thermocycling conditions varied betweenloci and that information is available from the authors upon re-quest. Relative allele sizes were determined automatically fromelectropherograms and then visually inspected to avoid errorscaused by automated calling.

Tests of assumptions and genetic diversity.—Loci weretested for null alleles, large allele dropout, and stutter peakswith MICROCHECKER version 2.2.3 (van Oosterhout et al.2004). An estimate of the microsatellite-scoring error rate wasgenerated by repeating 540 PCRs across 12 loci in 45 IGHindividuals, and the copy error rate per allele was calculatedas the ratio between the observed number of allelic differencesand total number of allelic comparisons (Bonin et al. 2004).Tests for linkage disequilibrium between all locus pairs ineach population and for conformance to Hardy–Weinbergproportions for each locus in each population, conducted usingthe Markov chain Monte Carlo approximation of Fisher’s exacttest, were implemented in GENEPOP version 4.0.10 (Rousset2008). Corrections for multiple tests used the Bonferronimethod (Rice 1989). Allelic richness (A), observed (Ho),and expected heterozygosity (He) were all determined withARLEQUIN version 3.1 (Excoffier et al. 2005). Standardizedprivate allelic richness (Ap) and standardized allelic richness(AR), equalized using rarefaction to a sample size of 38 genes,were calculated with HP-RARE version 1.0 (Kalinowski 2005).

Genetic structure.—A standardized measure of genetic dif-ferentiation, G ′

ST, (Hedrick 2005) was calculated between eachpopulation pair using FSTAT (Goudet 1995) in combinationwith RECODEDATA version 0.1 (Meirmans 2006). The G ′

STaccounts for differences in allelic diversity within populations,which can bias traditional FST estimators. Significance of ge-netic differentiation between population pairs, based on FST,was estimated using permutation tests implemented in FSTAT(Goudet 1995). Corrections for multiple tests used the FalseDiscovery Rate method (Benjamini and Yekutieli 2001), whichprovides increased power for detecting population divergencecompared with the Bonferroni method (Narum 2006). Signif-icance of isolation by distance, a correlation between genetic(G ′

ST) and geographic distances (river distances in kilometers),for all populations above the confluence of the Klamath and

Trinity rivers (see Results) was evaluated using a Mantel testin FSTAT (Goudet 1995; Hutchison and Templeton 1999) with10,000 permutations of the data. The geographic location forthe spring-run population at the Trinity River Hatchery wasestimated as 50 km upstream from the barrier dam, which is in-tended to reflect predam spawning activity (Moffett and Smith1950; Smith 1976).

Chord distances (Cavalli-Sforza and Edwards 1967) betweenpopulation pairs were calculated from the genetic data and usedto construct an unrooted neighbor-joining tree with PHYLIP ver-sion 3.68 (Felsenstein 1993). Branch support was evaluated us-ing 1,000 bootstrap replicates with branches appearing in morethan 90% of the trees considered well supported.

Population differentiation was graphically investigated byconducting a discriminant analysis of principal components(DAPC; Jombart et al. 2010) in the adegenet package (Jombart2008). The DAPC performs a preliminary data transformationstep using principal component analysis (PCA) to createuncorrelated variables that summarize total variability (e.g.,within and between groups). These variables are used asinput to discriminant analysis (DA), which aims to maximizebetween-group variability and achieve the best discriminationof multilocus genotypes into predefined clusters. Multivariateanalyses can provide clustering power comparable withBayesian methods but are free of assumptions about Hardy–Weinberg and linkage equilibria and take less time to calculate(Patterson et al. 2006; Jombart et al. 2010).

The Bayesian clustering algorithm employed in STRUC-TURE version 2.3.3 (Pritchard et al. 2000; Falush et al. 2003;Hubisz et al. 2009) was used to generate an ad hoc estimate ofthe most likely number of genetically distinct clusters present inthe data and an estimate of the proportion of each individual’sgenome assigned to each cluster. Each cluster representing a ge-netically distinct group is assumed to be free of Hardy–Weinbergand linkage disequilibrium. Estimates of the number of geneticclusters present in the data were generated by calculating thelog probability of the data [ln Pr(X|K)] and by estimating �K(Evanno et al. 2005), assuming the data consisted of K = 2, . . . ,13 genetically distinct groups. Twenty independent runs wereconducted at each value of K and runs that did not converge,as indicated by divergence distances among populations andlikelihoods, were discarded. STRUCTURE analyses were con-ducted with and without population locations as priors. To alignindependent STRUCTURE runs, we used the LargeKGreedyalgorithm (with 10,000 random input orders) as implementedin the software CLUMPP (Jakobsson and Rosenberg 2007).Graphical depictions of CLUMPP results were generated usingDISTRUCT (Rosenberg 2004).

RESULTS

Tests of Assumptions and Genetic DiversityOne locus (OTSG311) showed evidence of null alleles and

was removed from further analyses. None of the remaining 27

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 15:

05 1

4 M

ay 2

014

KLAMATH RIVER BASIN CHINOOK SALMON GENETIC STRUCTURE 1351

loci exhibited significant problems associated with null alleles,stutter peaks, or large allele dropout. Allelic scores were iden-tical in the 540 duplicate PCRs of fish from IGH, indicating avery low genotyping error rate. The estimated proportion of full-sibling pairs was 9% in TC, 2% in Horse Linto Creek (HLC),and 0% in all other populations. Tests of pairwise genetic dif-ferentiation (FST) between samples collected in different yearsbut from the same location indicated no significant differencesand thus were combined (Bonferroni corrected critical value =0.01). The South Fork Trinity River Spring (SFTS) sample wasexcluded from all analyses, except for the Bayesian cluster andmultivariate ones, due to small sample size (n = 7).

The loci were highly polymorphic, with an average of24.8 alleles per locus. Of the 351 tests for departure fromHardy–Weinberg proportions (13 populations at 27 loci), sevenwere significant after Bonferroni correction (critical value =0.00014). No single locus or population consistently departedfrom expectations, which indicated that locus- and population-specific factors were not causes for the observed departures. Atotal of 77 out of 4,563 tests for linkage disequilibrium were sig-nificant after Bonferroni correction (critical value = 0.000010).The majority of the significant tests (61) could be attributedto a single population, Salmon River Fall (SRF). This elevatedlinkage disequilibrium may be due to hybridization of this stockwith Salmon River Spring (SRS), Trinity River Hatchery (TRH),IGH, or a combination of those. The remaining significant testswere not characteristic of particular populations or locus pairs.Within-population measures of genetic diversity are presentedin Table 1.

Genetic StructureMost pairwise FST values differed significantly from zero

(critical value = 0.00321), and the main exception was thecomparisons between geographically proximate populations(Table 2). An initial test for isolation by distance indicated BC,TC, and HLC were outliers, probably because they belong toa different evolutionary lineage and ESU than the remaining

FIGURE 2. Relationship between pairwise genetic differentiation (G ′ST) and

river distance (RKM) for Klamath River Chinook Salmon populations abovethe confluence of the Klamath and Trinity rivers. Analysis excludes the HorseLinto Creek population.

populations (Gall et al. 1992). When these populations wereexcluded, the relationship between pairwise genetic differ-entiation (G ′

ST) and river distance was strong and significant(Mantel test: R2 = 0.80, P < 0.01; Figure 2). The intercept ofthe relationship was nearly zero (0.0002), which is consistentwith the classic definition of an isolation-by-distance model ofgene flow (Hutchison and Templeton 1999).

In the neighbor-joining tree, all nodes except two were sup-ported by bootstrap values greater than 90% (Figure 3). A groupfrom the lower basin (HLC, BC and TC) was resolved as dis-tinctive from the populations in the upper basin. The hatcherystocks from the Klamath (IGH) and Trinity rivers (TRH Spring

TABLE 2. Pairwise estimates of genetic differentiation (G ′ST) for Klamath River Chinook Salmon populations (below diagonal) and P-values for significance

tests of differentiation (above diagonal). Critical P-value adjusted for multiple tests using false discovery rate method was 0.00321. Significant tests are indicatedby an asterisk. See Table 1 for definition of population abbreviations. NA = data not available.

Population IGH BOG SHST SCOT SRS SRF TRHS TRHF SFTF HLC BC TC

IGH 0.37500 0.03141 0.00064* 0.00064* 0.00064* 0.00064* 0.00064* NA 0.00064* 0.00064* 0.00064*BOG −0.0112 0.76154 0.01218 0.00064* 0.00064* 0.00064* 0.00064* NA 0.00449 0.00064* 0.00192*SHST 0.0077 −0.0159 0.00064* 0.00064* 0.00064* 0.00064* 0.00064* NA 0.00064* 0.00064* 0.00064*SCOT 0.0280 0.0235 0.0238 0.00064* 0.00064* 0.00064* 0.00064* NA 0.00064* 0.00064* 0.00064*SRS 0.0741 0.0684 0.0685 0.0438 0.02885 0.00064* 0.00064* NA 0.00064* 0.00064* 0.00064*SRF 0.0560 0.0440 0.0536 0.0296 0.0096 0.00064* 0.00064* NA 0.00064* 0.00064* 0.00064*TRHS 0.1084 0.0892 0.1123 0.0885 0.0477 0.0354 0.00064* NA 0.00064* 0.00064* 0.00064*TRHF 0.0986 0.0903 0.0954 0.0604 0.0526 0.0424 0.0555 NA 0.00064* 0.00064* 0.00064*SFTF 0.0481 0.0512 0.0474 0.0235 0.0204 0.0274 0.0487 0.0184 NA NA NAHLC 0.1151 0.1267 0.1120 0.0817 0.0798 0.0743 0.1177 0.0921 0.0687 0.00064* 0.00128BC 0.1194 0.1228 0.1213 0.0864 0.0903 0.0806 0.1207 0.1143 0.0831 0.0565 0.00064*TC 0.2555 0.2397 0.2470 0.2225 0.1940 0.1833 0.2113 0.2413 0.2222 0.1832 0.1593

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 15:

05 1

4 M

ay 2

014

1352 KINZIGER ET AL.

FIGURE 3. Unrooted neighbor-joining tree of Klamath River basin ChinookSalmon populations. Branch lengths are proportional to chord distances andbootstrap support (>90%) is indicated along the branches. Population abbrevi-ations are defined in Table 1.

[TRHS] and TRH Fall [TRHF]) exhibited a large divergencefrom each other, and the other upper basin populations plottedin intermediate positions based upon their geographic distancefrom the main hatchery groups (Bogus Creek [BOG], ShastaRiver [SHST], Scott Creek [SCOT], South Fork Trinity RiverFall [SFTF], SRS, and SRF). Long terminal branches associatedwith TC are probably due to inclusion of siblings (see above).

In the DAPC, 90% of the total genetic variation was capturedby the first 205 principal components of PCA and these wereused as input to DA. The eigenvalues resulting from DA indi-cated that the first two axes were sufficient to summarize thegenetic structure of Klamath River Chinook Salmon (Figure 4,inset). The lower basin group (HLC, BC, and TC) diverged fromthe remaining populations along the first axis (Figure 4). Alongthe second axis, the largest divergence was between hatcherypopulations (IGH versus TRHS and TRHF), and the remain-ing tributary populations were situated along a line connectingthe hatchery populations. The coordinates of these populationsalong this line were related to their geographic distance fromthe hatcheries.

In the Bayesian cluster analysis, the ad hoc statistic �K indi-cated the strongest level of structure at K = 3 without the use oflocation information (Supplementary Figure 1 available in theonline version of this article). The three clusters included: (1)the lower basin (HLC, BC, and TC), (2) Klamath River (IGH,

BOG, and SHST), and (3) Trinity River (TRHS, TRHF, andSFTS) (Table 1; Figure 5). Populations SFTF, SCOT, SRS, andSRF were resolved as admixtures between the three primaryclusters. Inspection of individual admixture proportions at K =4 indicated a unique component primarily associated with mid-basin tributaries (SCOT, SRS, and SRF), and inspection of theplot at K = 5 indicated a division between TRHS and TRHF,which were resolved as distinct in a previous study (Kinzigeret al. 2008), supporting an assertion of at least K = 5 distinctclusters in our data. Even finer-scale structure was suggestedby inspection of outputs at K > 5, but these results should betreated with caution because Bayesian clustering methods canoverestimate the number of genetic clusters in data sets charac-terized by isolation by distance (Frantz et al. 2009; Kalinowski2011), as is the case here (see below). Bayesian cluster analysisusing location information provided similar patterns (results notshown).

DISCUSSION

Genetic StructureChinook Salmon from the Klamath River basin exhibit a

complex multilevel pattern of genetic structure defined primar-ily by geography. The three major genetic groups in the basinoriginate from separate geographic regions: the lower basin, theKlamath River, and the Trinity River. This result is supportedby concordant patterns in neighbor-joining trees, multivariateplots, and Bayesian model-based clustering, all of which indi-cated divergence between populations from these regions. Theimportance of geography in describing genetic structure in Kla-math River Chinook Salmon is consistent with large-scale phy-logeographic assessments that indicate geography explains themajority of genetic structure within this species (Waples et al.2004; Beacham et al. 2006; Seeb et al. 2007; Moran et al. 2013).

A distinct lower basin group of Chinook Salmon was iden-tified, including populations in Blue, Terwer, and Horse Lintocreeks. This finding is consistent with previous studies thatdesignated Chinook Salmon populations from below the conflu-ence of the Klamath and Trinity rivers into the Southern Oregonand Northern California Coast ESU (Figure 1; Gall et al. 1992;Myers et al. 1998). Populations below the confluence are moreclosely related to Chinook Salmon from other basins in northernCalifornia and southern Oregon than to stocks above the Trinityand Klamath river confluence (Gall et al. 1992; Myers et al.1998). A similar pattern has been observed for Klamath Riverbasin steelhead O. mykiss populations (Pearse et al. 2007).However, contrary to previous investigations (Gall et al. 1992),our analysis identified Horse Linto Creek as part of the lowerbasin group, despite its location upstream of the ESU boundaryat the confluence of the Klamath and Trinity rivers (Myers et al.1998). This result suggests a shift in ESU affiliation between1987 and 1997, the dates of field collections for the previousand present studies. Horse Linto Creek has a relatively smallpopulation of Chinook Salmon, and it is possible that annual

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 15:

05 1

4 M

ay 2

014

KLAMATH RIVER BASIN CHINOOK SALMON GENETIC STRUCTURE 1353

FIGURE 4. Scatterplot of the first two principal components of DAPC using population locations as prior clusters. Populations are labeled inside their 95%inertia ellipsis and dots represent individuals. The inset indicates the eigenvalues of the first 12 principal components. Population SFTF superimposes SFTS andSRF superimposes SRS. Population abbreviations are defined in Table 1. [Figure available online in color.]

influxes of migrants from larger populations cause changes ingenetic composition of this population that results in temporalvariability in phylogeographic affiliation. However, if the affin-ity of this population with the lower basin group remains stable,a reconsideration of the ESU boundary may be warranted.

The Upper Klamath and Trinity Rivers ESU was resolved asa genetically distinct lineage, but composed of two divergentsubgroups, identified as the Klamath and Trinity herein. Thesetwo groups were found to be connected by processes that can bedescribed by a one-dimensional, linear, stepping-stone modelwhere gene exchange occurred primarily, but not exclusively,between adjacent populations (Kimura and Weiss 1964). The

boundaries of the model are defined by Iron Gate Hatchery andTrinity River Hatchery, which exhibited the highest levels ofgenetic differentiation in pairwise tests and large separation inmultivariate plots, and are geographically located at the terminiof anadromy (Table 2; Figures 1, 4). Spawning populations ingeographic areas between the hatcheries, such as the Shasta,Scott, Salmon, and South Fork Trinity rivers, were at interme-diate positions in tree-based and multivariate analyses and theircoordinates were related to geographic distance from the hatch-ery populations. The strong relationship between genetic andgeographic distance further attests to the importance of riverdistance in mediating gene flow among populations (Figure 2).

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 15:

05 1

4 M

ay 2

014

1354 KINZIGER ET AL.

FIGURE 5. Individual membership coefficients for Klamath River ChinookSalmon for K = 3–5 clusters without the use of population location informationfrom Bayesian cluster analysis. Concordance of the alignments across the inde-pendent STRUCTURE runs (H′) was 0.99, 0.87, and 0.79 for K = 3, 4, and 5,respectively. Each individual is represented by a single horizontal bar dividedin K colored segments, with lengths that are proportional to cluster membershipassignments. Population abbreviations are defined in Table 1. [Figure availableonline in color.]

Our proposed gene flow model for the Klamath and Trin-ity groups is concordant with the linear arrangement of pop-ulations along the river course, strong natal homing behaviorof Chinook Salmon (Quinn 1993; Dittman and Quinn 1996),and predictions about gene flow based on tag recoveries ofhatchery-origin salmon. The proportions of naturally spawningfish found to be of Iron Gate Hatchery origin in tributaries situ-ated 0 (Bogus Creek), 21 (Shasta River), and 75 river kilometers(RKM) (Scott River) downstream from the hatchery were 0.33,0.12, and 0.00, respectively (California Department of Fish andWildlife, unpublished data). Long-distance straying from IronGate Hatchery to Trinity River Hatchery (400 RKM apart) is

rare and migrants between them number only a few individualseach year (California Department of Fish and Wildlife, unpub-lished data). While both genetic and field data have inherentproblems for estimating the extent of migration among popula-tions (Koenig et al. 1996; Whitlock and McCauley 1999), bothsupport the contention that gene exchange decreases with riverdistance and occurs primarily between adjacent populations.Other populations of anadromous salmonids, including steel-head in the Klamath River, conform to isolation-by-distancemodels of gene flow (Hendry et al. 2004; Primmer et al. 2006;Palstra et al. 2007; Pearse et al. 2007; Narum et al. 2010), andour inferred pattern for Klamath River Chinook Salmon is con-sistent with patterns from these other closely related species.

Genetic structure was also evident among populations fromthe same subbasin that exhibit temporally distinct migrationtiming, although divergence between these populations was lessthan genetic differences observed between stocks from sepa-rate geographic regions. We examined three sympatric spring-and fall-run pairs from the Trinity River Hatchery, South ForkTrinity River, and Salmon River. Pairwise genetic differentia-tion was low but significant between runs from the Trinity RiverHatchery, nonsignificant between Salmon River runs, and sam-ple sizes were insufficient for comparison of South Fork TrinityRiver runs. We selected samples from the early portion of thespring-run spawning period and the end of the fall-run spawningperiod for the Trinity River stock. Employing a similar samplingstrategy in the Salmon and South Fork Trinity rivers may pro-vide improved power for resolving potential genetic differencesbetween spring- and fall-run salmon in these tributaries. Model-based clustering analysis further supported Trinity River Hatch-ery runs as distinct groups, but not the sympatric pairs from othertributaries (Figure 5). Previous work on Trinity River HatcheryChinook Salmon has demonstrated hybridization and gradualtransition in genetic composition between spring- and fall-runsalmon throughout the spawning season (Kinziger et al. 2008).Here, we found that spring- and fall-run fish from the same trib-utary of the Klamath River basin were genetically most similarto one another and divergent from other groups in the basin(Figures 3, 4), which indicates that ecotypic variation appearsto have evolved independently through a process of parallel evo-lution, which is relatively common in Chinook Salmon (Wapleset al. 2004; Beacham et al. 2006; Moran et al. 2013).

Midbasin Group and Hatchery StockingBayesian cluster analysis resolved an additional genetic

group that primarily contained individuals from midbasin Kla-math River tributaries (Scott River and Salmon River spring-and fall-run stocks). Individual assignment coefficients indi-cated substantial introgression of the midbasin group by theTrinity and Klamath hatchery stocks (Figure 5). The geographicintermediacy of this group between the Iron Gate and Trin-ity River hatcheries also makes this a logical location for ad-mixture (Figure 1). The cluster analysis also suggested strongasymmetry in introgression between the hatchery stocks and the

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 15:

05 1

4 M

ay 2

014

KLAMATH RIVER BASIN CHINOOK SALMON GENETIC STRUCTURE 1355

midbasin group. Population admixture coefficients indicated thehatchery stocks comprised 40% of the Salmon River ancestry,whereas the midbasin group comprised only 10% of the hatch-ery stocks ancestry. While it is possible that these patterns mayreflect historic gene flow patterns in the basin, we hypothesizethat asymmetric introgression is a result of construction of im-passable dams and associated mitigation hatcheries, which haveconsistently released millions of fish annually since the early1960s. The dramatically larger number of fish produced in theseprograms, relative to any natural population, should result inhigher immigration pressure from the hatchery programs intonatural populations than vice versa. While hatchery-origin fishare almost never observed in the Salmon River, populations lo-cated immediately below both hatcheries in the upper reachesof both the Trinity and Klamath rivers receive enough migrantsto result in complete integration with the hatchery stocks (seeabove). While the effects of the hatchery stocks attenuate withdistance, we hypothesize that migrants from populations belowthe hatcheries and dams serve as stepping stones across multi-generation time scales for extending introgressive influence todownstream Chinook Salmon populations outside the direct in-fluence of hatchery stocks.

Supplementation Impacts on Genetic StructureDespite a history of large-scale hatchery supplementation

spanning more than half a century, Klamath River basin Chi-nook Salmon exhibited significant genetic differentiation amongpopulations and retained genetic structure patterns similar tothat predicted by an equilibrium between migration and ge-netic drift, as well as that observed more broadly in the species(Hendry et al. 2004; Waples et al. 2004; Beacham et al. 2006;Moran et al. 2013). This stands in stark contrast to populationstructure of California Central Valley Chinook Salmon, wherefall-run populations have been homogenized as a consequenceof hatchery activities (Williamson and May 2005). While ourresults indicate introgressive pressure from the hatchery stocksinto natural populations, they also suggest that the geographicscale of the Klamath River basin is large enough to have al-lowed retention of significant genetic structure among KlamathRiver Chinook Salmon populations despite such supplementa-tion. Nevertheless, it is possible that current patterns are unstableand do not reflect historic conditions prior to initiation of large-scale hatchery programs. Comparison of historical populationgenetic structure to contemporary patterns, through analysis ofhistorical tissue collections, would provide a conclusive evalua-tion of the extent to which hatchery operations and habitat lossfrom impassable dams has modified Chinook Salmon geneticpopulation structure in the Klamath River basin.

ACKNOWLEDGMENTSFunding for this project came from the Hoopa Valley Tribe

and the National Marine Fisheries Service, Southwest FisheriesScience Center. We thank A. Antonetti, S. Borok, P. Brucker,

B. Chesney, D. Hillemeier, G. Kautsky, E. Logan, B. Matilton,N. Pennington, and A. Sprowles for providing tissue samplesused in this study, and A. Abadıa-Cardoso, V. Apkenas, D.Faulkner, E. Gilbert-Horvath, and H. Starks for assistance withlaboratory data collection. Reference to trade names does notimply endorsement by the U.S. Government.

REFERENCESAllendorf, F. W., and S. R. Phelps. 1981. Use of allelic frequencies to describe

population structure. Canadian Journal of Fisheries and Aquatic Sciences38:1507–1514.

Banks, M. A., V. K. Rashbrook, M. J. Calavetta, C. A. Dean, and D. Hedgecock.2000. Analysis of microsatellite DNA resolves genetic structure and diversityof Chinook Salmon (Oncorhynchus tschawytscha) in California’s CentralValley. Canadian Journal of Fisheries and Aquatic Sciences 57:915–927.

Beacham, T. D., K. L. Jonsen, J. Supernault, M. Wetklo, L. Deng, and N.Varnavskaya. 2006. Pacific Rim population structure of Chinook Salmonas determined from microsatellite analysis. Transactions of the AmericanFisheries Society 135:1604–1621.

Benjamini, Y., and D. Yekutieli. 2001. The control of the false discovery rate inmultiple testing under dependency. Annals of Statistics 29:1165–1188.

Bjorkstedt, E. P., B. C. Spence, J. C. Garza, D. G. Hankin, D. Fuller, W. E.Jones, J. J. Smith, and R. Macedo. 2005. An analysis of historical popula-tion structure for evolutionarily significant units of Chinook Salmon, CohoSalmon, and steelhead in the north-central California coast recovery domain.NOAA Technical Memorandum NMFS-SWFSC-382.

Bonin, A., E. Bellemain, P. Bronken Eidesen, F. Pompanon, C. Brochmann, andP. Taberlet. 2004. How to track and assess genotyping errors in populationgenetics studies. Molecular Ecology 13:3261–3273.

Cavalli-Sforza, L. L., and A. W. F. Edwards. 1967. Phylogenetic analysis: modelsand estimation procedures. American Journal of Human Genetics 19:233–257.

Dittman, A. H., and T. P. Quinn. 1996. Homing in Pacific salmon: mechanismsand ecological basis. Journal of Experimental Biology 199:83–91.

Evanno, G., S. Regnaut, and J. Goudet. 2005. Detecting the number of clusters ofindividuals using the software STRUCTURE: a simulation study. MolecularEcology 14:2611–2620.

Excoffier, L., G. Laval, and S. Schneider. 2005. Arlequin (version 3.0): an inte-grated software package for population genetics data analysis. EvolutionaryBioinformatics Online 1:47–50.

Falush, D., M. Stephens, and J. K. Pritchard. 2003. Inference of populationstructure using multilocus genotype data: linked loci and correlated allelefrequencies. Genetics 164:1567–1587.

Felsenstein, J. 1993. PHYLIP (phylogeny inference package), version 3.68.University of Washington, Seattle.

Finnegan, A. K., and J. R. Stevens. 2008. Assessing the long-term geneticimpact of historical stocking events on contemporary populations of AtlanticSalmon, Salmo salar. Fisheries Management and Ecology 15:315–326.

Frantz, A. C., S. Cellina, A. Krier, L. Schley, and T. Burke. 2009. Using spa-tial Bayesian methods to determine the genetic structure of a continuouslydistributed population: clusters or isolation by distance? Journal of AppliedEcology 46:493–505.

Gall, G. A. E., D. Bartley, B. Bentley, J. Brodziak, R. Gomulkiewicz, andM. Mangel. 1992. Geographic variation in population genetic structure ofChinook Salmon from California and Oregon. U.S. National Marine FisheriesService Fishery Bulletin 90:77–100.

Garza, J. C., S. M. Blankenship, C. Lemaire, and G. Charrier. 2008. Geneticpopulation structure of Chinook Salmon (Oncorhynchus tshawytscha) in Cal-ifornia’s Central Valley. CALFED Bay-Delta Program, Final Report, Sacra-mento, California.

Goudet, J. 1995. FSTAT (version 1.2): a computer program to calculate F-statistics. Journal of Heredity 86:485–486.

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 15:

05 1

4 M

ay 2

014

1356 KINZIGER ET AL.

Hamilton, J., G. L. Curtis, S. M. Snedaker, and D. K. White. 2005. Distributionof anadromous fishes in the upper Klamath River watershed prior to hy-dropower dams—a synthesis of the historical evidence. Fisheries 30(4):10–20.

Hamilton J., D. Rondorf, M. Hampton, R. Quinones, J. Simondet, and T. Smith.2011. Synthesis of the effects to fish species of two management scenariosfor the secretarial determination on removal of the lower four dams on theKlamath River. U.S. Fish and Wildlife Service, Yreka, California.

Hansen, M. M. 2002. Estimating the long-term effects of stocking domesti-cated trout into wild Brown Trout (Salmo trutta) populations: an approachusing microsatellite DNA analysis of historical and contemporary samples.Molecular Ecology 11:1003–1015.

Hansen, M. M., D. J. Fraser, K. Meier, and K. L. D. Mensberg. 2009. Sixty yearsof anthropogenic pressure: a spatio-temporal genetic analysis of Brown Troutpopulations subject to stocking and population declines. Molecular Ecology18:2549–2562.

Hansen, M. M., E. E. Nielsen, D. Bekkevold, and K. L. D. Mensberg. 2001.Admixture analysis and stocking impact assessment in Brown Trout (Salmotrutta), estimated with incomplete baseline data. Canadian Journal of Fish-eries and Aquatic Sciences 58:1853–1860.

Hansen, M. M., E. E. Nielsen, and K. L. D. Mensberg. 1997. The problem ofsampling families rather than populations: relatedness among individuals insamples of juvenile Brown Trout Salmo trutta L. Molecular Ecology 6:469–474.

Healey, M. C. 1991. Life history of Chinook Salmon (Oncorhynchustshawytscha). Pages 311–394 in C. Groot and L. Margolis, editors. Pacificsalmon life histories. University of British Columbia Press, Vancouver.

Hedrick, P. W. 2005. A standardized genetic differentiation measure. Evolution59:1633–1638.

Hendry, A. P., V. Castric, M. T. Kinnison, and T. P. Quinn. 2004. The evolutionof philopatry and dispersal: homing versus straying in salmonids. Pages 52–91 in A. P. Hendry and S. C. Stearns, editors. Evolution illuminated: salmonand their relatives. Oxford University Press, New York.

Hubisz, M. J., D. Falush, M. Stephens, and J. K. Pritchard. 2009. Inferringweak population structure with the assistance of sample group information.Molecular Ecology Resources 9:1322–1332.

Hutchison, D. W., and A. R. Templeton. 1999. Correlation of pairwise geneticand geographic distance measures: inferring the relative influences of geneflow and drift on the distribution of genetic variability. Evolution 53:1898–1914.

Jakobsson, M., and N. A. Rosenberg. 2007. CLUMPP: a cluster matching andpermutation program for dealing with label switching and multimodality inanalysis of population structure. Bioinformatics 23:1801–1806.

Jombart, T. 2008. adegenet: a R package for the multivariate analysis of geneticmarkers. Bioinformatics 24:1403–1405.

Jombart, T., S. Devillard, and F. Balloux. 2010. Discriminant analysis of prin-cipal components: a new method for the analysis of genetically structuredpopulations. BMC Genetics [online serial] 11:article 94.

Kalinowski, S. T. 2005. HP-RARE 1.0: a computer program for performing rar-efaction on measures of allelic richness. Molecular Ecology Notes 5:187–189.

Kalinowski, S. T. 2011. The computer program STRUCTURE does not reliablyidentify the main genetic clusters within species: simulations and implicationsfor human population structure. Heredity 106:625–632.

Kalinowski, S. T., A. P. Wagner, and M. L. Taper. 2006. ML-RELATE: a computerprogram for maximum likelihood estimation of relatedness and relationship.Molecular Ecology Notes 6:576–579.

Kimura, M., and G. H. Weiss. 1964. The stepping stone model of populationstructure and the decrease of genetic correlation with distance. Genetics49:561–576.

Kinziger, A. P., E. J. Loudenslager, D. G. Hankin, E. C. Anderson, andJ. C. Garza. 2008. Hybridization between spring- and fall-run ChinookSalmon returning to the Trinity River, California. North American Journal ofFisheries Management 28:1426–1438.

Koenig, W. D., D. Van Vuren, and P. N. Hooge. 1996. Detectability, philopatry,and the distribution of dispersal distances in vertebrates. Trends in Ecologyand Evolution 11:514–517.

Martinez, J. L., J. Dumas, E. Beall, and E. Garcia-Vazquez. 2001. Assessingintrogression of foreign strains in wild Atlantic Salmon populations: variationin microsatellites assessed in historic scale collections. Freshwater Biology46:835–844.

Matala, A. P., S. R. Narum, W. Young, and J. L. Vogel. 2012. Influences of hatch-ery supplementation, spawner distribution, and habitat on genetic structureof Chinook Salmon in the South Fork Salmon River, Idaho. North AmericanJournal of Fisheries Management 32:346–359.

Meirmans, P. G. 2006. Using the AMOVA framework to estimate a standardizedgenetic differentiation measure. Evolution 60:2399–2402.

Moffett, J. W., and S. E. Smith. 1950. Biological investigations of the fisheryresources of Trinity River, California. U.S. Fish and Wildlife Service SpecialScientific Report—Fisheries 12.

Moran, P., D. J. Teel, M. A. Banks, T. D. Beacham, M. R. Bellinger, S. M.Blankenship, J. R. Candy, J. C. Garza, J. E. Hess, S. R. Narum, L. W.Seeb, W. D. Templin, C. G. Wallace, and C. T. Smith. 2013. Divergent life-history races do not represent Chinook Salmon coast-wide: the importance ofscale in Quaternary biogeography. Canadian Journal of Fisheries and AquaticSciences 70:415–435.

Myers, J. M., R. G. Kope, G. J. Bryant, D. Teel, L. J. Lierheimer, T. C.Wainwright, W. S. Grant, F. W. Waknitz, K. Neely, S. T. Lindley, and R.S. Waples. 1998. Status review of Chinook Salmon from Washington, Idaho,Oregon, and California. NOAA Technical Memorandum NMFS-NWFSC-35.

Narum, S. R. 2006. Beyond Bonferroni: less conservative analyses for conser-vation genetics. Conservation Genetics 7:783–787.

Narum, S. R., J. E. Hess, and A. P. Matala. 2010. Examining genetic lineages ofChinook Salmon in the Columbia River basin. Transactions of the AmericanFisheries Society 139:1465–1477.

Palstra, F. P., M. F. O’Connell, and D. E. Ruzzante. 2007. Population structureand gene flow reversals in Atlantic Salmon (Salmo salar) over contemporaryand long-term temporal scales: effects of population size and life history.Molecular Ecology 16:4504–4522.

Patterson, N., A. L. Price, and D. Reich. 2006. Population structure andeigenanalysis. PLoS (Public Library of Science) Genetics [online serial]2(12):e190.

Pearse, D. E., C. J. Donohoe, and J. C. Garza. 2007. Population genetics of steel-head (Oncorhynchus mykiss) in the Klamath River. Environmental Biologyof Fishes 80:377–387.

Primmer, C. R., A. J. Veselov, A. Zubchenko, A. Poututkin, I. Bakhmet, andM. T. Koskinen. 2006. Isolation by distance within a river system: geneticpopulation structuring of Atlantic Salmon, Salmo salar, in tributaries of theVarzuga River in northwest Russia. Molecular Ecology 15:653–666.

Pritchard, J. K., M. Stephens, and P. Donnelly. 2000. Inference of populationstructure using multilocus genotype data. Genetics 155:945–959.

Quinn, T. P. 1993. A review of homing and straying of wild and hatchery-produced salmon. Fisheries Research 18:29–44.

Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43:223–225.Rosenberg, N. A. 2004. DISTRUCT: a program for the graphical display of popu-

lation structure. Molecular Ecology Notes 4:137–138.Rousset, F. 2008. GENEPOP’007: a complete re-implementation of the GENEPOP

software for Windows and Linux. Molecular Ecology Resources 8:103–106.Seeb, L. W., A. Antonovich, M. A. Banks, T. D. Beacham, M. R. Bellinger,

S. M. Blankenship, M. R. Campbell, N. A. Decovich, J. C. Garza, C. M.Guthrie III, T. A. Lundrigan, P. Moran, S. R. Narum, J. J. Stephenson, K.T. Supernault, D. J. Teel, W. D. Templin, J. K. Wenburg, S. F. Young, andC. T. Smith. 2007. Development of a standardized DNA database for ChinookSalmon. Fisheries 32:540–552.

Smith, F. E. 1976. Water development impact on fish resources and associatedvalues of the Trinity River, California. Pages 98–111 in J. F. Orsborn andC. H. Allman, editors. Instream flow needs, volume 2. American FisheriesSociety, Western Division, Bethesda, Maryland.

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 15:

05 1

4 M

ay 2

014

KLAMATH RIVER BASIN CHINOOK SALMON GENETIC STRUCTURE 1357

Snyder, J. O. 1931. Salmon of the Klamath River California. California Depart-ment of Fish and Game, Fish Bulletin 34.

Spence, B. C., W. G. Duffy, J. C. Garza, B. C. Harvey, S. M. Sogard, L. A.Weitkamp, T. H. Williams, and D. A. Boughton. 2011. Historical occurrenceof Coho Salmon (Oncorhynchus kisutch) in streams of the Santa Cruz moun-tain region of California: response to an endangered species act petition todelist Coho Salmon south of San Francisco Bay. NOAA Technical Memo-randum NMFS-SWFSC-472.

Van Oosterhout, C., W. F. Hutchinson, D. P. M. Wills, and P. Ship-ley. 2004. MICRO-CHECKER: software for identifying and correcting geno-

typing errors in microsatellite data. Molecular Ecology Notes 4:535–538.

Waples, R. S., D. J. Teel, J. M. Myers, and A. R. Marshall. 2004. Life-historydivergence in Chinook Salmon: historic contingency and parallel evolution.Evolution 58:386–403.

Whitlock, M. C., and D. E. McCauley. 1999. Indirect measures of gene flowand migration: FST �= 1/(4Nm + 1). Heredity 82:117–125.

Williamson, K. S., and B. May. 2005. Homogenization of fall-run ChinookSalmon gene pools in the Central Valley of California, USA. North AmericanJournal of Fisheries Management 25:993–1009.

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 15:

05 1

4 M

ay 2

014